A Comparative Algorithmic Stratification to Identify Impact of Brain Tumor on Substantia Nigra in Humans

Brain is a pivotal part of the human body that requires accurate diagnosis and analysis on a periodical basis. The human brain is segmented into various regions, with the Substantia Nigra (SN) playing a decisive part in controlling the motor functions in the human body. The impact that brain tumors hold in terms of their size and location is an imperative attribute that determines the efficient functioning of substantia nigra. The existing research is distributed largely to comprehend the disfunctionalities in the brain, and analysis of motor inefficiencies, but fails to bridge the correlative breach that tumor impact has on substantia nigra. This research pivots on the understudied aspect of substantia nigra in relation to the density of the brain tumor through a juxtaposed algorithmic approach. The proposed indagation is effectuated with Binary Generalized Linear Model (GLM) Logistic Regression, Neural Network and an Optimizable Ensemble model effectuated by combining AdaBoost, Decision Tree, Neural Network, Binary GLM Logistic regression and Support vector machine. The simulations explicitly delineated that the optimizable ensemble delivered higher accuracy of classifying the tumors impacting Substantia Nigra, as compared to the other algorithmic approaches that explicated lower stratification accuracy. The simulations are implemented in MATLAB, and the results are successfully obtained.